Confessions of a crass Keynesian

The German federal minister of finance, Peer Steinbrueck, does not like anything that increases government deficits.  He does not like them, Sam-I-Am.  I believe he is wrong – very wrong and dangerously wrong. In the interest of Anglo-German harmony and ever-closer cooperation, I have written this post.

It explains that there are bad deficits and good deficits.  Or, in the words of Ecclesiastes: “To every thing there is a season, and a time to every purpose under the heaven:” a time to cut taxes and a time to raise taxes, a time to borrow and a time to refrain from borrowing.

Today is a time, even Ecclesiastes would agree, made for increased government borrowing, provided a few key conditions are satisfied.

Mum, the government are running a deficit again!

Government deficits have to be financed by selling assets, by borrowing from the domestic private sector, from the rest of the world or from the central bank.  Asset sales by the government can be ignored as a financing option for the governments of the USA, the UK and the nations that constitute the Euro Area.  Quite the opposite has been happening lately, with governments acquiring large stake in domestic banks and other financial institutions.

Borrowing from the central bank

Borrowing from the central bank (selling Treasury debt to the central bank) requires the central bank either to increase its monetary liabilities (currency or bank reserves held with the central bank), or to increase its non-monetary liabilities, or to run down its stock of official foreign exchange reserves or other central bank assets.  The USA, the Euro Area and the UK all have floating exchange rates, and foreign exchange market intervention has  not been a significant pastime for the monetary authorities of these countries.  Under current economic circumstances, financing the acquisition of additional Treasury debt by running down central bank holdings of private securities would not make sense.  True non-monetary liabilities of the central bank (Central Bank Bills or Central Bank Bonds) are common in developing countries and emerging markets but have not been a common sight on the balance sheets of the Fed, the ECB and the Bank of England – until recently.

The Federal Reserve System today holds a large amount (more than $400 bn) of Treasury deposits on its balance sheet , the result of the Treasury selling Treasury securities to the public and depositing the money with the Fed. The Fed has used these funds to acquire additional private securities.  Instead of borrowing from the Treasury (through the Treasury’s deposits with the Federal Reserve System), the Fed is currently considering the possibility of issuing non-monetary, interest-bearing securities directly to the market.  Assuming Treasury and Fed securities of the same maturity are perfect substitutes for private investors, this would give the Fed another, economically equivalent mechanism  for expanding its balance sheet without increasing the monetary base.  Why the Fed would want to increase the size of its balance sheet by issuing non-monetary liabilities rather than monetary liabilities is unclear to me.  Liquidity preference remains unbounded from above.  Further quantitative easing is not inflationary for as long as current economic conditions last.

Once the official policy rate is at its zero floor, quantitative and qualitative easing are the main instruments of monetary policy, which becomes inextricably intertwined with liquidity management.  By acquiring longer-dated government securities and financing these purchases by expanding the base money stock, central banks can bring down the risk-free nominal rate of interest at longer maturities than overnight.  Such purchases of longer-maturity government securities reinforce the  expectations mechanism – long-term risk-free nominal yields are tied to current expectations of future overnight rates, give or take a term premium.  By acquiring private securities, including illiquid private securities, whether through outright purchase or as collateral in repos or at the discount window, the central bank can influence a range of term spreads and liquidity spreads on these private securities.

For simplicity and to put the issue as sharply as possible, let’s assume that when the government (the Treasury) borrows from the central bank, the central bank monetises its acquisition of the Treasury securities, that is,  it increases the sum of currency and banks’ deposits (reserves) with the central bank. In practice, the increase in base money is likely to take the form mainly of larger bank reserves with the central bank.

As long as the economy is in the doldrums, with a large (or even large and growing) amount of spare capacity and extreme risk-averse behaviour of banks, other financial institutions and individual investors, the increased quantity of central bank money will be absorbed willingly at the current price level and at the current (near zero) level of the short-run nominal interest rate.  Fear and loathing in the financial markets have created a near unbounded liquidity preference – a  willingness to hold a humongous quantity of real base money.  Such injections of base money are therefore not inflationary.

When the economy recovers, as it will, and private investors recover their bottle, the demand for real base money normalises and the private sector finds itself with excessive real base money balances at the current official policy rate and price level.  The private sector will try to reduce its holdings of real money base money balances partly by switching their portfolio allocation towards non-monetary assets and partly by spending them.  In the aggregate, of course, the private sector cannot reduce the nominal stock of base money, unless the central bank plays ball and de-monetises the public debt it had monetised earlier. If it does not do so, monetary equilibrium will have to restored through a higher general price level.

This de-monetisation of the public debt (the reversal of the earlier monetisation) will be automatic if, when the economy recovers, the official policy rate rises again above its zero lower bound and quantitative easing comes to an end.  When the official policy rate is set (pegged) above its lower bound, the demand for real base money balances becomes finite again.  With the general price level pre-determined (given/sticky in the short run because the world is crass-Keynesian in the short run, that is, in real time), the nominal base money stock becomes endogenous.  Given the central bank’s balance sheet, the counterpart of the endogenous (and lower) stock of base money is the endogenous (and lower) stock of Treasury securities held by the central bank.

When the economy normalises, the public debt issued by the Treasury to finance any deficits incurred during the slump leaves the central bank and comes back home to mama.  Mama will have to convince the markets (the domestic private sector and/or the rest of the world) that it wants to hold this public debt. If the interest rates at which the markets are willing to hold that debt are high, or if there is no interest rate level, however high, at which the markets wish to hold the additional public debt spewed out of the central bank’s balance sheet, we have a problem.  Either the government forces the central bank to hang on to the Treasury debt or the economy will have to live with very high interest rates or, in the most extreme case, with default on the public debt.

The first scenario – permanent monetisation of public debt issuance – means that, when the economy recovers, the central bank is forced to engage in whatever amount of monetary issuance may be required to finance the government deficit.  The result will be inflation, when the economy recovers – quite possibly inflation in excess of the explicit or implicit inflation target of the central bank

While it is therefore true that the government can always, if it has the power to tell the central bank what to do, monetise the outstanding stock of government debt and any amount of new issuance of government debt, no matter how large, as long as the debt is denominated in domestic currency, there is a limit, for most base money demand functions, to the amount of real resources the government can extract though the inflation tax.  This implies that there is a limit to the real value of the government deficit that can be financed through the inflation tax and also to the amount of index-linked government debt and foreign-currency-denominated government debt that can be monetized and inflated away.

The nastiest alternative is that the real value of the government deficit is larger than the real value of the additional issuance of money balances that the private sector is willing to absorb at any constant rate of inflation: the maximum long-run inflation tax at a constant rate of inflation is less than the real value of the government deficit.  In that case hyperinflation will result.

It is a long way from the current threat of deflation (negative inflation) to hyperinflation, but it is never to soon to start worrying about the next crisis.

Borrowing from the market

Now consider the case where the government deficit is financed by borrowing from the markets rather than from the central bank.  Like every other economic agent, the government is subject to an intertemporal budget constraint.  A government is solvent if the value of its net stock of outstanding debt does not exceed the present discounted value of its current and future primary surpluses.  The government’s primary surplus is its conventional financial surplus plus net interest paid on its outstanding stock of debt. Since the government here excludes the central bank, among the government revenues that are included in the government’s primary surplus are the taxes paid by the central bank to the Treasury.  These contributions of the central bank to the government budget are not usually referred to as ‘taxes’.

Central bank operating profits (net interest income and other income minus the cost of running the show) are usually split between a contribution paid into the government budget and an addition to the central bank’s reserves. The contribution of the central bank to the government budget (called taxes on the central bank in the previous paragraph) increase one-for-one with any increase in interest paid by the government to the central bank on the central bank’s holdings of government securities.  At the margin, therefore, borrowing from the central bank is free to the government.

As regards the solvency of the central bank, it makes no difference whether base money is non-interest-bearing (the case of currency) or interest bearing (often the case with banks’ reserves with the central bank).  Ultimately, the central bank can settle any domestic-currency denominated claim on itself by paying in currency, which is both non-interest-bearing and irredeemable.

When the government violates its ex-ante intertemporal budget constraint or solvency constraint (its outstanding debt is larger than the present discounted value of its planned/expected primary surpluses) there are but three options for closing this ‘solvency gap’.  (1) it cuts current and/or future public spending; (2) it raises current and/or future tax revenues; or (3) it defaults on part or all of the sovereign debt.

When will the future spending cuts or tax increases have to be implemented?  The solvency constraint and intertemporal budget constraint are silent on this matter.  They only assert that the present discounted value of current and future spending cuts and tax increases has to be at least equal to the solvency gap.  It does not tell you when this has to happen.  So could we wait until the years 3125 before spending is cut or taxes are increased?  Market realities imply the answer is no.  Markets are doubting Thomases.  To them seeing is believing.  They want to put their fingers in the wounds.  In practice, spending will have to be cut and/or taxes will have to be increased as soon as this is sensible from a conjectural or cyclical point of view.  As soon as a tax increase or public spending cut would be counter-cyclical rather than pro-cyclical, it will have to be implemented.  Failure to do so at the first opportunity would weaken the credibility of the government.  Markets will entertain steadily stronger doubts about the sustainability of the fiscal-financial programme of the government.  Default risk premia will be added to the interest rates at which the government borrows.  As the perceived likelihood of a sovereign default increases, the default risk premia will rise and, ultimately, the government will be rationed out of the primary debt markets: it will become impossible to add to the government’s net indebtedness and even to roll over maturing debt.

So is Steinbrueck right in condemning proposals for deficit-financed fiscal stimuli in Europe and elsewhere to counteract the contraction of effective private demand?  This question has two parts: (1) does a temporary tax cut or spending increase, followed by a future tax increase or spending cut that restores government solvency stimulated demand?  Can governments credibly commit themselves to raise future taxes or cut future public spending by enough to maintain government solvency if they deliver an immediate tax cut or public spending increase?

Does a temporary tax cut boost consumer spending?

For the moment, let’s assume that the answer to the second question is ‘yes’ and let’s address the first.  I will focus on a temporary tax cut.  Will a temporary tax cut today resulting in a larger budget deficit and increased government borrowing stimulate demand, if future government taxes are raised again by the same amount, in present discounted value, as the current tax cut?  Or, in other words, does postponing taxes, holding constant their present discounted value, boost demand?

I will focus on cuts in household taxes, like the personal income tax or VAT.  The argument that deficit-financed tax cuts don’t boost consumption demand is known as Ricardian equivalence or debt neutrality.  For it to be true, the aggregate consumption demand of consumers has to behave in the same way as would the consumption of a representative infinite-lived consumer with perfect foresight.  This consumer knows, when his taxes are cut, that he will pay higher taxes in the future and that the present value of current and future taxes has not changed.  His permanent income or wealth have not changed.  He will not feel better off as the result of the tax cut.  He will save all of the tax cut to pay the higher future taxes.

The demographics of the Ricardian equivalence model are not convincing.  People are born, live for a while and die.  While they are alive, they overlap with earlier generations (the old) and with generations born since their own generation arrived (the young).  Postponing taxes will therefore shift the burden of paying the taxes from the older generations to the younger generations, and possibly even to the (as yet) unborn.  The usual life-cycle arguments suggest that the old (who have fewer remaining years to live) will have a higher marginal propensity to consume out of a temporary tax cut than the young.  The old certainly will have a higher marginal propensity to consume than the unborn.  So cutting taxes today and raising them again in the future by the same amount in present discounted value raises aggregate consumption demand.

It is important that the current tax cut and the future tax increase don’t affect the same people equally in both periods.  For the fiscal stimulus to work through a life-cycle mechanism, the current tax cut would primarily have to benefit today’s old and working generations.  The future tax increase would be paid mainly by today’s young and working generations, or by those who today are still unborn (future generations). This will be the case if the tax is a tax on labour income or a capitation or head tax.  It would not be true if the future tax increase were on the income from an asset that is already in existence and fully owned today (land or physical capital).  In that case, both the current tax cut and the future tax increase will be reflected in the value of the assets, which will not change.  Taxes on future labour income are not, however, capitalised in the value of any asset owned by anyone currently alive.  This is because we have abolished hereditary slavery: the human capital of future generations is not owned by anyone currently alive today.  Postponing labour income taxes therefore redistributes resources from the young and the unborn to the old.  The life-cycle For life-cycle reasons, the old have a higher marginal propensity to consume than the old, and the unborn don’t consume at all.

If current generations care about their descendants, they may be planning to leave bequests for them.  Should the government then try, by cutting taxes today and raising them in the future, to redistribute towards parents and grandparents and away from their children and their grand children, the parents and the grand parents would simply offset this involuntary intergenerational redistribution by the government with voluntary intergenerational redistribution towards their descendants.  Lower taxes today would be saved and left as increased bequests.  Since most people don’t leave bequests in the first place (most retirement wealth is annuitized), this ingenious argument in favour of Ricardian equivalence even in a world of overlapping generations with finite life spans, is a theoretical curiosum, not a useful empirical benchmark.

In addition to life-cycle reasons for current tax cuts boosting aggregate consumption, there are liquidity reasons.  If some consumers are liquidity-constrained (unable to borrow more or sell assets) a cut in current taxes will relax a binding liquidity constraint on current spending, even if the consumer were to be fully aware that he would have to pay higher taxes in the future.  Of course, not all households can be liquidity-constrained, otherwise there would be no-one to purchase the securities the government is issuing to finance the increased government deficit.

In the current liquidity crunch there is bound to be a significant increase in the number of liquidity-constrained households.  If they could be targeted through the tax cuts, the consumption effects would be strengthened.  Liquidity constraints are especially likely among those with large debts, no liquid assets and no collateralisable assets who suffer a temporary interruption in their labour income, due to unemployment, say.  They are also likely to affect those with rising age-earnings profiles who have few liquid and collateralisable assets.  This would include yuppies and other upwardly mobile groups.

It is hard to believe that, provided a government has the fiscal-financial credibility to be able to commit itself to future tax increases or public spending cuts when it implements immediate tax cuts or public spending increases, that this would fail to stimulate aggregate demand through the usual life-cycle effects and liquidity constraint effects.

The VAT cut rubbished so emphatically by the German minister of finance is in fact quite a clever tax cut, precisely because it is temporary.  By cutting the price to the consumer today and raising it again tomorrow, there is an incentive to shift the timing of consumption of non-durables and services, and the timing of the purchases of consumer durables, toward the present, when consumer prices are temporarily low.  The neo-classical substitution effect reinforces the Keynesian current disposable income effects.

Mr. Steinbrueck is not impressed and provides variations on the ‘who would cross the road for a 2.5 percent VAT cut when there are 20 percent to 50 percent discounted sales on everywhere’ argument.  I think Mr. Steinbrueck underestimates the German and British shopper.  But even if he were right and the substitution effect of the temporary VAT cut is negligible, there still is the income effect.

Would the income effect have been stronger if, instead of a VAT cut worth, say £14 bn, the same amount of money had been sent directly to British households in the form of a cheque with the same amount of money for each tax-paying or benefit-receiving adult?  This is not at all obvious to me.  Assume households spend the same amount following the VAT cut as they did before.  If prices come down by the full 2.5 percent cut in the VAT rate, they will buy a larger amount of real commodities with the same amount of income.  This stimulates the demand for real goods and services.  If prices were to come down by less than the cut in VAT, after-tax profits would increase for the sellers, which could boost the consumption demand of their owners or the demand for investment or working capital inputs by the enterprises themselves.

If none of this sounds convincing to the German minister of finance, he could always implement a temporary investment credit or a similar temporary subsidy to or tax cut on investment on fixed assets. Precisely because it is temporary, it would shift the timing of investment spending toward the present.

So Mr. Steinbrueck’s outburst appears to be rooted in faulty logic and sloppy thinking.

Who can afford even a temporary tax cut or spending increase?

Until further notice, I will assume in what follows that the central banks in the countries or monetary union I am discussing stick to their price stability or dual price stability and full employment mandates.  That means that they will monetise government debt and deficits only up to the point where they perceive such actions to undermine the effective pursuit of price stability.

Not all nations in the north Atlantic region are equally well positioned to implement a fiscal stimulus that would result in a significant increase in the government deficit.  The decision of the EU to call on all EU member states to implement a 1.5 percent of GDP stimulus to GDP therefore appears to be ill-advised.  The magnitude of the stimulus should be modulated (a) according to the needs of the country (how deep is the recession, how open is the economy) and (b) according to the fiscal-financial sustainability of the government and the credibility of the government, that is, the likelihood that it will act in a determined counter-cyclical manner during the next economic upswing, raising taxes and/or cutting public spending.

Italy’s fiscal-financial sustainability and the credibility of its government were it to announce a pleasure today – pain tomorrow temporary fiscal stimulus are close to zero.  The UK’s fiscal-financial sustainability is poor and the credibility of its government is severely impaired after years of pro-cyclical fiscal policy during the age of excess that preceded the current bust.  The UK government, by de facto or de jure underwriting the liabilities of the UK banking system has assumed debts worth over 400 percent of GDP.  Of course there are assets on the other side of the banks’ balance sheets, but the liabilities are firm and clear, while the assets are dodgy and of uncertain value.  The same applies to the United States of America, where the Federal government is not only up to its neck in actual and contingent liabilities through its underwriting of the banking system, GSEs like Fannie Mae and Freddie Mac, insurance companies like AIG and non-specific partly financial enterprises like GE, but is about to have the water rise even higher as it bails out the three domestic automobile manufacturers.

Germany’s Maastricht gross general government debt as a percentage of annual GDP was about 20 percentage points higher than that of the UK at the end of 2007.  However, the cyclically adjusted budget deficit in Germany is far smaller than that of the UK.  In addition, the exposure of the German government to its banking sector, while non-trivial, is much smaller than that of the UK government to its over-developed banking sector.  Most important, the German authorities have demonstrated both the willingness and the capacity to engage in countercyclical fiscal policy during the most recent boom period.

This means that reasons of national self-interest and as a constructive member of the global community, Germany can and should engage in a significantly larger fiscal stimulus (relative to the size of its economy) than the US and the UK.  Spain and France also should deliver an above-average fiscal stimulus, while Italy cannot afford much of a stimulus at all.

Sovereign default versus inflation levies

For the first time since the German default of 1948, a number of countries in the north Atlantic region (North America and Western Europe) face a non-negligible risk of sovereign default.  The main driver is their governments’ de facto or de jure underwriting of the balance sheets of their banking sectors and, in some cases, of a range of non-bank financial and non-financial institutions deemed too big to fail.  Unfortunately, in a number of cases, the aggregate of the institutions deemed too large, too interconnected or too politically connected to fail may also be too large to save.  The solvency gap of the private institutions the authorities wish to save exceeds the fiscal spare capacity of the sovereign.

The clearest example of the ‘too large to save’ problem is Iceland.  Iceland’s government did not have the fiscal resources to bail out their largest three internationally active banks.  The outcome was that all banks went into insolvency.  The government then nationalised some key domestic parts of the three banks out of the insolvency regime, decided (under massive pressure from the British, Dutch and German governments) to honour Iceland’s deposit guarantees and left the rest of the unsecured debt to be resolved through the insolvency process.

Other countries face the problem of the inconsistent quartet ((1) a small open economy; (2) a large internationally exposed banking sector; (3) a national currency that is not a major international reserve currency; and (4) limited fiscal capacity).  They include Switzerland, Sweden, Denmark and the UK.  Ireland, the Netherlands, Belgium and Luxembourg have all but the third of these characteristics.

There can be little doubt that, faced with the choice between sovereign default and an unexpected burst of inflation to reduce the real value of the government’s domestic-currency-denominated debt, the US government would choose inflation.  It would simply instruct the Fed to produce the required burst of inflation.  The Fed is the least independent of the leading central banks.  The Fed regained a measure of operational independence in the conduct of monetary policy in 1951 through the US Treasury Federal Reserve Accord.  This accord does not have the force of law, and can be revoked at any time by the Treasury.

In the UK too, I believe that, given the choice between sovereign default and a burst of unanticipated inflation, the UK Treasury would choose inflation.  The Treasury could repatriate the rate setting powers of the Monetary Policy Committee of the Bank of England under the Reserve Powers clause of the Bank of England Act 1998.

Things are different in the Euro Area.  The independence of the ECB is embedded in the Treaties.  A unanimous decision by all member states is required to change the Treaty.  Given this operational independence ‘on steroids’ of the ECB, it is unlikely that any Euro Area national government or coalition of governments could bully the ECB into engaging in a burst of public-debt-busting unanticipated inflation.  Perhaps Mr Peer Steinbrueck’s intemperate expostulations about the horrors of increased public debt are due to his recognition that he, unlike his fellow ministers of finance in the UK and the US, does not have the option of inflating away the public debt, unless Germany were to decide to leave the Euro Area.

If instead we accept as an axiom that every German finance minister worth his salt would emulate the stance taken by Ludwig Erhard in 1948 and would therefore never choose the inflation option, even if the only alternative would be government default, then Peer Steinbrueck’s eruption is hard to rationalise.  Perhaps it cannot be rationalised because it was an emotional outburst rather than a thought-through argument.  Surely not….

Maverecon: Willem Buiter

Willem Buiter's blog ran until December 2009. This blog is no longer active but it remains open as an archive.

Professor of European Political Economy, London School of Economics and Political Science; former chief economist of the EBRD, former external member of the MPC; adviser to international organisations, governments, central banks and private financial institutions.

Willem Buiter's website